Almost all modern datetime libraries support precision up to nanoseconds. There has been some discussion in the past to add it to Python, but there seems to be limited demand and it wouldn't fit into the current API.
Since whenever breaks from the standard library, it makes sense to use nanoseconds, if only for consistency.
There is probably no performance drawback since both fit into a 32-bit integer
Almost all modern datetime libraries support precision up to nanoseconds. There has been some discussion in the past to add it to Python, but there seems to be limited demand and it wouldn't fit into the current API.
Since whenever breaks from the standard library, it makes sense to use nanoseconds, if only for consistency. There is probably no performance drawback since both fit into a 32-bit integer